Long-Lasting Sound-Evoked Afterdischarge in the Auditory Midbrain

Abstract

Different forms of plasticity are known to play a critical role in the processing of information about sound. Here, we report a novel neural plastic response in the inferior colliculus, an auditory center in the midbrain of the auditory pathway. A vigorous, long-lasting sound-evoked afterdischarge (LSA) is seen in a subpopulation of both glutamatergic and GABAergic neurons in the central nucleus of the inferior colliculus of normal hearing mice. These neurons were identified with single unit recordings and optogenetics in vivo. The LSA can continue for up to several minutes after the offset of the sound. LSA is induced by long-lasting, or repetitive short-duration, innocuous sounds. Neurons with LSA showed less adaptation than the neurons without LSA. The mechanisms that cause this neural behavior are unknown but may be a function of intrinsic mechanisms or the microcircuitry of the inferior colliculus. Since LSA produces long-lasting firing in the absence of sound, it may be relevant to temporary or chronic tinnitus or to some other aftereffect of long-duration sound.

Figure 1. GABAergic and nonGABAergic neurons in IC were distinguished in vivo.

(a) GAD67, ChR2-YFP, and MAP2 were immunohistochemically stained. Arrows show the neurons in which GAD67 and ChR2-YFP are colocalized. Asterisks indicate the non-GABAergic neurons with only MAP2. (b) GAD67 and ChR2-YFP are colocalized in axonal terminals. Scale = 10 μm. (c) A neuron activated by light. Left, the spike responses to light (30 ms, 50 mW). Right, response to 200 ms white noise (80 dB). Gray box indicates the sound presentation here and in subsequent figures. (d) A neuron suppressed by light. Left upper, the response to 200 ms white noise (80 dB). Left lower, the white noise with light (30 ms, 50 mW). Blue box indicates light presentation. The right panel is the raster plot of the response to noise and light. (e) An image of a juxtacellularly stained neuron that was suppressed by light. Note that it is GAD67 (green signal, a marker for GABA) negative. All the stained neurons with light suppression were GAD67 negative (n = 5).

Figure 2. Both GABAergic and nonGABAergic neurons in IC had LSA.

(a,c) The response of neurons with LSA to sound. (a) A nonGABAergic neuron. (c) A GABAergic neuron. Upper trace is the voltage spike trace of the response and the lower trace is its PSTH. The gray traces are magnified PSTHs. The arrows indicate the dip (D) and peak (P) of the PSTHs. The Gray boxes indicate the sound presentation. The sound intensity was 70dB in (a) and 80 dB in (c). (b,d) Pseudocolor map shows the number of spikes within 30 s after the sound termination as RDS firing rate and sound duration were varied. (b,d) are the responses from the neuron in (a,c), respectively. (e) The response of a GABAergic neuron without LSA. The sound intensity was 70 dB. (f) Histograms of the minimum sound duration that evoked LSA in each neuron. (g) The number of LSA spikes was plotted against the number of RDS spikes. Different symbols indicate the responses from different neurons. Black and red symbols indicate the responses from GABAergic and nonGABAergic neurons, respectively, in this and subsequent figures. In the inset, the responses from CBA/J mice are indicated by green filled squares, circles, and triangles. (h,i) The peak time (h) and decay (i) of LSA PSTH are plotted against the number of LSA spikes. In the insets, the responses from CBA/J mice are indicated by green filled squares, circles, and triangles. (j) The comparison of the accommodation of RDS to 30 s (left) and 60 s (right) sound between the LSA+ and LSA− neurons. The larger index indicates less adaptive response in RDS. Black and red boxplots and dots indicate the response of GABAergic neurons and nonGABAergic neurons, respectively.

Figure 3. An overview of the responses of LSA+ neurons.

(a) Pseudocolor maps of the number of spikes within 20 s after sound termination as shown in Fig. 2b,d. (G1-G4) GABAergic neurons. (N1-N5) nonGABAergic neurons. (CBA1-CBA3) the recordings from CBA/J mice. Pseudocolor maps were made when we successfully recorded the responses to more than three test sounds. The maps shown in Fig. 2 are excluded here. (b) The distribution of the LSA+ neurons in IC in VGAT-ChR2 mice. Black and red circles indicate GABAergic and nonGABAergic neurons, respectively. The location of each neuron was plotted on the schematic sections. Thin lines indicate the boundary of central nucleus of IC (ICC). ICC was separated in three different regions by the ratio of GABAergic and glycinergic terminals. In region 1, glycinergic terminals dominate. In region 2, there are relatively more GABAergic terminals than glycinergic terminals. In the most dorsal region (3) GABAergic terminals predominate although a few glycinergic terminals are still present. The numbers beside the circles correspond to the numbers in Fig. 3a. G0 and N0 correspond to the cells in Fig. 2d,b, respectively.

Figure 4. Discontinuous sound evoked LSA.

(a) The response of a GABAergic neuron to sound with 50% duty cycle of 1 s (50 times repetition). Upper trace is the voltage spike trace of the response and the middle trace is its PSTH. The lower traces are magnified voltage traces. (b,c) The plots of RDS (left) and interstimulus spikes (right) against time. (b) LSA+. (c) LSA−.